摘要
针对红外图像对比度和分辩率低、噪声大的特点,提出一种基于平稳小波域的红外图像增强新方法。对高频子带中幅值较小的噪声系数进行衰减,幅值较大的边缘细节系数进行放大;对低频子带系数的幅值采用所提出的正弦函数进行伸缩处理。实验结果表明,本文提出的方法在有效地增强红外图像对比度及边缘细节的同时,又能很好地抑制背景噪声,综合性能明显优于传统的直方图均衡化和反锐化掩膜增强方法。
According to the low contrast,low resolution and big noise characteristics of the infrared image, a new meth- od for infrared image enhancement based on stationary wavelet domain is proposed. The high-frequency sub-band noise coefficients of the smaller amplitude are attenuated, while the bigger edge detail coefficients are magnified. The amplitude of the low-frequency sub-band coefficients are scale enhance processed by the proposed sine function. The experimental results show that the proposed method enhances the image contrast and edge detail of the infrared image, and also suppresses background noise very well. Comprehensive performance is better than traditional histogram equalization and unsharp mask enhancement method.
出处
《激光与红外》
CAS
CSCD
北大核心
2013年第6期703-707,共5页
Laser & Infrared
基金
广东省自然科学基金项目(No.S2012010010368)资助
关键词
红外图像
图像增强
平稳小波变换
正弦变换
infrared image
image enhancement
stationary wavelet transform
sine transform
